A Real-Time Subband SAR Imaging Algorithm Based on an Approximate Echo Signal Model

Conventional synthetic aperture radar (SAR) imaging algorithms establish the echo signal model based on the exact hyperbolic range equation and compensate for the range-azimuth coupling through complicated approximations to obtain high-resolution SAR images. Consequently, they incur high hardware requirements and implementation costs for real-time SAR imaging. This paper proposes an approximate SAR echo signal model based on two-dimensional block partitioning. Within each data block, the hyperbolic range equation is fitted with a straight line. Provided that a given accuracy requirement is met, the range-azimuth coupling inside the data block is neglected, thereby simplifying the range cell migration correction (RCMC). Based on this model, the two-dimensional matrix-form processing is decomposed into two one-dimensional vector-form processing steps. Through subband division, azimuth compression is decomposed into multiple short fast Fourier transform (FFT) operations at reduced sampling rates. The proposed algorithm reduces the computational load and storage requirements while ensuring imaging performance, thereby lowering the resource demands for real-time SAR imaging. After introducing the working principle and imaging model of SAR, this paper establishes an approximate SAR echo model based on two-dimensional block partitioning and analyzes its approximation errors; presents a subband imaging workflow; discusses the algorithm’s computational complexity, memory requirements, and valid parameter range; and verifies the algorithm’s effectiveness by processing both point target simulation data and real data.